AM Best’s July 30, 2026 market segment report on rated captive insurers shows net premiums written growing 65.5% over the five years through 2025, with a 7% increase in 2025 alone. AM Best rates more than 220 captives globally, roughly 70% of them domiciled in the United States. The report also flags combined ratio pressure across that rated universe for two consecutive years and estimates $8.2 billion in savings generated for sponsoring organizations over the same five-year window.
The combined ratio pressure is the signal that matters for reserve adequacy. Its source is consistent with what happens when a program built on favorable reserve releases from smaller prior accident years runs out of runway.
Who it affects
Single-parent captive owners and captive boards whose written premium has grown more than 30% over the past five years face this problem directly. Risk retention groups and group captives where membership or covered payroll has expanded materially are in the same position. If the program is smaller and growth has been concentrated in recent years, the credibility problem is worse: the older accident years in the development triangle represent a fundamentally different book.
The reserve mechanism
Chain-ladder development and Bornhuetter-Ferguson (BF) both depend on a stable, consistent triangle. Chain-ladder derives age-to-age development factors from prior accident years. When those years represent 40% or 50% of the exposure volume of the current year, the factors they produce are anchored to a smaller program’s experience: fewer large claims, lower aggregate severity, and in some lines shorter development lags.
The BF method blends captive development history with an a priori expected loss ratio. If that a priori is drawn from recent captive experience rather than an external industry benchmark, it inherits the same credibility gap. Favorable prior-year development releases, which buffered reported loss ratios during the growth period, get embedded in the a priori as if they reflect the ongoing program rather than a smaller prior vintage.
Two consecutive years of combined ratio pressure at rated captives is consistent with that buffer running dry. As prior favorable development fades, the underlying loss experience on the current, larger book surfaces without the offset.
The correction is straightforward but requires raising the question explicitly. Actuaries can calibrate tail factors against external development benchmarks from NCCI, ISO, or CAS industry triangles matched to the captive’s lines. BF a priori expected loss ratios can be anchored to those same external benchmarks rather than to the captive’s own recent, favorably buffered history. See IBNR for single-parent captives for how these two methods interact in practice, and captive funding at a confidence level for why the credibility gap also affects confidence interval analysis and confidence-level funding decisions.
Actuarial Standards Board ASOP 43 requires actuaries to identify and discuss material sources of uncertainty in reserve estimates. Rapid exposure growth is one of the named sources. An actuarial opinion that does not address whether external benchmarks were used to supplement a credibility-limited internal triangle may not fully satisfy the standard.
What this means for your next review
Ask your actuary one concrete question: what fraction of the tail factors in the current reserve opinion are derived from the captive’s own development history, and what is the ratio of exposure in the oldest accident year to exposure in the most recent year?
If that ratio is below 60%, the tail factors are calibrated to a materially smaller program. Push for external benchmarks as a credibility check. For a captive that has grown significantly, funding at the 50th percentile using a triangle anchored to smaller prior years may be the functional equivalent of funding below the 40th percentile against the current book.